Differentiation is a process whereby structures and functions of cells are progressively committed to give rise to more specialised cells, such as the formation of T cells or B cells from immature haemopoietic precursors. Therefore, as the cells become more committed, they become more specialised. In the majority of mammalian cell types, cell differentiation is a one-way process leading ultimately to terminally differentiated cells. However, although some cell types persist throughout life without dividing and without being replaced, many cell types do continue to divide during the lifetime of the organism and undergo renewal. This may be by simple division (e.g. liver cells) or, as in the case of cells such as haemopoietic cells and epidermal cells, by division of relatively undifferentiated stem cells followed by commitment of one of the daughter cells to a programme of subsequent irreversible differentiation. All of these processes, however, have one feature in common: cells either maintain their state of differentiation or become more differentiated. They do not become undifferentiated or even less differentiated.
Retrodifferentiation is a process whereby structures and functions of cells are progressively changed to give rise to less specialised cells. Some cells naturally undergo limited reverse differentiation (retrodifferentiation) in vivo in response to tissue damage. For example, liver cells have been observed to revert to an enzyme expression pattern similar to the foetal enzymic pattern during liver regeneration (Curtin and Snell, 1983, Br. J. Cancer, Vol 48; 495-505).
Jose Uriel (Cancer Research, 1976, vol 36, pp 4269-4275) presented a review on the topic of retrodifferentiation, in which he said:                “retrodifferentiation appears as a common adaptive process for the maintenance of cell integrity against deleterious agents of varied etiology (physical, chemical, and viral). While preserving the entire information encoded on its genome, cells undergoing retrodifferentiation lose morphological and functional complexity by virtue of a process of self-deletion of cytoplasmic structures and the transition to a more juvenile pattern of gene expression. This results in a progressive uniformization of originally distinct cell phenotypes and to a decrease of responsiveness to regulatory signals operational in adult cells. Retrodifferentiation is normally counterbalanced by a process of reontogeny that tends to restore the terminal phenotypes where the reversion started. This explains why retrodifferentiation remains invariably associated to cell regeneration and tissue repair.”        
Uriel (ibid) then went on to discuss cases of reported retrodifferentiation—such as the work of Gurdon relating to nuclei from gut epithelial cells of Xenopus tadpoles (Advances in Morphogenesis, 1966, vol 4, pp 1-43. New York Academic Press, Eds Abercrombie and Bracher), and the work of Bresnick relating to regeneration of liver (Methods in Cancer Research, 1971, vol 6, pp 347-391).
Uriel (ibid) also reported on work relating to isolated liver parenchymal cells for in vitro cultures. According to Uriel:                “Contrary to the results with fetal or neonatal hepatocytes, with hepatocytes from regenerating liver, or from established hepatomas, it has been difficult to obtain permanent class lines from resting adult hepatocytes.”        
Uriel (ibid) also reported on apparent retrodifferentiation in cancer, wherein he stated:                “the biochemical phenotypes of many tumours show analogous changes of reversion toward immaturity . . . during the preneoplastic phase of liver carcinogenesis, cells also retrodifferentiate.”        
More recent findings on retrodifferentiation include the work of Minoru Fukunda (Cancer Research, 1981, vol 41: 4621-4628). Fukunda induced specific changes in the cell surface glycoprotein profile of K562 human leukaemic cells by use of the tumour-promoting phorbol ester, 12-O-tetradecanoyl-phorbol-13-acetate (TPA). According to Fukunda TPA appeared to induce the K562 human leukaemic cells into a retrodifferentiated stage.
Also, Hass et al. (Cell Growth & Differentiation, 1991, vol 2: 541-548) reported that long term culture of TPA-differentiated U-937 leukaemia cells in the absence of phorbol ester for 32-36 days resulted in a process of retrodifferentiation and that the retrodifferentiated cells detached from the substrate and reinitiated proliferation.
As mentioned above, another reported case of retrodifferentiation is the work of Curtin and Snell (Br. J. Cancer, 1983, vol 48: 495-505). These workers compared enzymatic changes occurring during diethylnitrosamine-induced hepatocarcinogenesis and liver regeneration after partial hepatectomy to normal liver differentiation. Theses workers found changes in enzyme activities during carcinogenesis that were similar to a step-wise reversal of differentiation. According to these workers, their results suggest that an underlying retrodifferentiation process is common to both the process of hepatocarcinogenesis and liver regeneration.
More recently, Chastre et al. (FEBS Letters, 1985, vol 188 (2), pp 2810-2811) reported on the retrodifferentiation of the human colonic cancerous subclone HT29-18.
Even more recently, Kobayashi et al. (Leukaemia Research, 1994, 18 (12): 929-933) have reported on the establishment of a retrodifferentiated cell line (RD-1) from a single rat myelomonocyticleukemia cell which differentiated into a macrophage-like cell by treatment with lipopolysaccharide (LPS).
Much of the above prior art focuses on retrodifferentiation as a stage in carcinogenesis. Several prior art documents refer to experiments where tumour cell lines have apparently been retrodifferentiated. However, these prior art experiments were carried out using tumour cell lines. The situation in genetically aberrant tumour cell lines is not comparable with the normal differentiation pathways. Indeed, it is questionable whether these results indicate true retrodifferentiation in the sense of normal cell lineages. Further, the vast majority of the prior art retrodifferentiated cells were incapable of redifferentiating to a more committed cell, whether of the same lineage, or of any other lineage. One exception is given in Kobayashi et al., 1994, Vol 18; 929-933 where a retrodifferentiated tumour cell line was differentiated into a macrophage-like cell using lipopolysaccharide. However, the retrodifferentiation achieved was very limited and the cells remained committed both before and after treatment.
Similarly, the reverse differentiation seen to occur naturally in liver cells is also very limited and can more accurately be classed as modulations of the differentiated state, that is to say, reversible changes between closely related cell phenotypes.